pyAGI is a Python package that functions as an autonomous agent for your Python application. By interfacing with OpenAI models, it utilizes a sequence of prompts to instruct the AI in generating various outputs such as application descriptions, architecture outlines, user experience flows, and even code samples. All these are achieved using the langchain framework to construct chains of language models.
- Supports several OpenAI models, including a range of GPT-3 and Davinci models.
- Provides functionality for generating:
- Application descriptions
- Software architecture outlines
- User experience flows
- Code flow descriptions
- Coding steps
- Application code
- Application descriptions
To use pyAGI, you need to have an OpenAI account and obtain your API key. Once you have your OpenAI API key, set it as an environment variable OPENAI_API_KEY in your system.
export OPENAI_API_KEY='your-api-key-here'
To install pyAGI, use pip:
pip install pyAGI
The package is available at PyPI pyAGI (https://pypi.org/project/pyAGI/).
After installing, you can use pyAGI to generate content for your application as follows:
from pyAGI.pyagi import PyAGI
Create a PyAGI instance with a selected OpenAI model
py_agi = PyAGI.create_llm_chain(selected_model="text-davinci-003")
# Run the PyAGI instance with the objective
py_agi({"objective": "Your Objective", "selected_model": "text-davinci-003"})
python main.py "Objective of your app" "OpenAI Model"
For example:
python main.py "Build a weather app" "text-davinci-003"
- "text-davinci-002",
- "text-davinci-003",
- "gpt-3.5-turbo",
- "gpt-3.5-turbo-0301",
- "text-curie-001",
- "text-babbage-001",
- "text-ada-001"
Upcoming models (https://openai.com/waitlist/gpt-4-api):
- "gpt-4",
- "gpt-4-0314",
- "gpt-4-32k",
- "gpt-4-32k-0314"
Please note that upcoming models will not be supported until released.
Contributions are welcome! Please open an issue if you encounter a bug, or a pull request if you have improvements to the existing code.
This project is licensed under the MIT License.